Europe’s Defence Startup Boom: $8.7B Raised and the Dual-Use Regulation Trap

The €8 Billion Moment Europe Did Not Plan For

Five years ago, European venture capital had effectively blacklisted defence technology. The ethical consensus in the tech investment community held that backing autonomous weapons, surveillance systems, or military AI was a reputational liability that no fund return could justify. That consensus collapsed — first with Russia’s invasion of Ukraine in 2022, then with a series of NATO members formally committing to spend 2% of GDP on defence, and finally with the economic returns becoming too large to ignore.

By 2025, European defence, security, and resilience (DSR) startups raised a record $8.7 billion, according to a joint report by Dealroom and the NATO Innovation Fund published in February 2026. That figure is a 55% increase year-over-year and nearly four times higher than 2020 levels. Late-stage mega-rounds tripled to $4.7 billion, signalling that the ecosystem has moved from early-stage experimentation to serious scaling. AI dominated the sector, underpinning 44% of all DSR funding — the highest share in six years — as autonomous systems, defence-optimised computing, and AI-enabled decision-making became central to next-generation military capability.

The UK led Europe with $2.9 billion raised in 2025 ($9.9 billion since 2020), with Germany closing the gap at $2.1 billion. Munich alone has attracted $7 billion in DSR investment — making it the fastest-growing hub in Europe. Sofia and Oslo appeared in the top-5 European DSR cities for the first time in five years, reflecting the geographic broadening of the ecosystem.

In early 2026, the deals continued. Harmattan AI raised $200 million in a Series B round for its Paris-based autonomous combat systems, backed by Dassault Aviation. TYTAN Technologies closed €30 million for its Munich-based autonomous interceptor drones, co-led by Armira and the NATO Innovation Fund. Frankenburg Technologies raised €30 million for Estonian missile production. Kelluu secured €15 million for Finnish autonomous airships for NATO surveillance — supported by the NATO Innovation Fund after a review of 1,300 applications from which 44 startups were selected.

The Regulatory Wall That Capital Cannot Buy Through

Here is what the capital surge does not fix: a founder in Amsterdam whose counter-drone technology is funded by European institutional money, validated by a NATO member’s military, and ready to sell — can still be legally blocked from exporting to another NATO ally without navigating 25+ national export-licence systems that do not speak to each other.

The EU’s dual-use export control framework — Regulation (EU) 2021/821, updated in September 2025 to expand controls over quantum computers, advanced microchips, and AI systems that can be used for both civilian and military purposes — governs which technologies require a licence before they can be sold or transferred outside the EU. The September 2025 update was one of the most substantial revisions in recent years, broadening the scope of controlled items to include cryogenic-temperature electronics and parametric signal amplifiers alongside the already-regulated areas of cryptography, sensors, and navigation systems.

For a defence startup, the practical consequences are severe. Onodrim Industries, the Amsterdam-based sensing and defence manufacturing company that raised €40 million at seed stage in early 2026, must evaluate the export-control status of every component in its technology stack before signing a sales agreement with a non-EU buyer. A drone with an AI navigation system may trigger three separate control categories simultaneously. And that is before reaching the import-country’s own licensing requirements, which vary by NATO member and add another compliance layer.

DEPO Ventures General Partner Petr Šíma, whose firm bridges Central and Eastern European talent with Western capital, put it plainly: “When founders first meet us, they often don’t know if their technology falls under strict defence or dual-use regimes.” That uncertainty is not a compliance detail — it is a valuation question. Buyers conduct due diligence on export-licence risk; a startup without clean export-control documentation is worth less to a potential acquirer or a public-contract customer. Compliance has become as critical as product innovation for valuation and investor confidence.

The European Commission’s AGILE pilot scheme, introduced in spring 2026 with €115 million in funding through 2027, attempts to address part of the problem by issuing small grants within a four-month approval window to firms working on urgent operational needs. But AGILE addresses the procurement cycle on the government side, not the export-control compliance burden on the startup side. Founders who receive an AGILE grant still need to build the legal architecture that allows them to sell the resulting technology outside their home market.

What Founders Should Do About the Dual-Use Trap

1. Classify Your Technology Stack Before You Take the First VC Cheque

The moment a defence startup accepts institutional investment, its technology stack becomes a compliance asset or liability. Every component — the AI model, the sensor package, the navigation system, the communications protocol — should be mapped against the EU Dual-Use Regulation Annex I and the Wassenaar Arrangement’s munitions and dual-use lists. This is not a post-Series A legal project; it is a pre-seed architecture decision. Founders who build with controlled components without knowing they are controlled will face either a costly redesign or a permanently constrained addressable market. Get a dual-use export-control attorney involved at the term-sheet stage, not at the Series B.

2. Architect for Modular Compliance, Not Monolithic Products

A defence product built as a monolithic system — where the AI layer, the autonomous navigation layer, and the communication layer are inseparable — creates a single export-control classification problem that affects the entire product. A modular architecture, where the controlled components can be separated from the uncontrolled ones, allows the startup to sell the civilian applications of its technology without triggering military-export licence requirements. This is not a theoretical design choice: Harmattan AI’s decision to build its combat system on an open-architecture platform with modular subsystems was, in part, driven by the need to maintain both defence and civil aviation sales tracks simultaneously.

3. Build the End-User Control Process as a Sales Infrastructure Asset

One of the most common compliance failures for European DSR startups — identified in regulatory reviews and VC due-diligence reports — is the absence of end-user control processes. An end-user control process is the documentation trail that proves who bought your technology, for what stated purpose, and with what legal commitments against re-export or diversion. Without this infrastructure, a startup cannot legally export to many markets, cannot satisfy NATO Innovation Fund portfolio reporting requirements, and cannot demonstrate to an acquirer that its customer base is not a regulatory liability. Build the end-user control process as a sales infrastructure asset from the first contract, not as a remediation project before an acquisition.

The Structural Question: Can Europe Build a Coherent Defence Innovation Ecosystem?

The record $8.7 billion raised by European DSR startups in 2025 is, paradoxically, partly a symptom of the fragmentation it is trying to solve. Capital concentrates in Munich, London, and Berlin — not because those cities have the best founders, but because they have the regulatory familiarity, the legal infrastructure, and the defence-ministry relationships that founders need to navigate the compliance environment. A founder in Tallinn building missile-production software for the Estonian military — exactly the kind of domestic defence capability NATO members are urgently trying to build — raises her next round in London because that is where the VC ecosystem that understands her technology’s export-licence status is located.

The NATO Innovation Fund’s model — reviewing 1,300 applications to select 44 startups — creates an elite filter that produces funded, validated companies, but it does not change the underlying 25-national-system fragmentation that those companies must still navigate after the cheque arrives. The AGILE scheme’s €115 million is too small to solve a structural problem that requires either a single European export-control authority or a mutual-recognition framework between member states. That institutional reform is the missing piece that venture capital cannot fund.

For now, the opportunity is real and the constraint is manageable with the right legal architecture. A European DSR founder who understands dual-use classification, builds modular products, and constructs end-user control processes from day one can access the $8.7-billion market that has formed — and position their company for the institutional reform that will eventually arrive.

Frequently Asked Questions

Sources & Further Reading

• European Defence, Security and Resilience Startups Raise $8.7B — NATO Innovation Fund / Dealroom
• Defence Startups Draw €8bn to Europe’s Hottest Tech Investment Spot — EU Perspectives
• Europe’s Top Defence Deals in Early 2026 — Vestbee
• Defense Tech Under Legal Scrutiny: How Regulation Is Reshaping Startup Investments in Europe — The Recursive
• EU Updates Dual-Use Export Control List: Key Changes for Emerging Technologies — Akin Gump
• EU Dual-Use Export Controls — EUR-Lex